Tropical forage breeding from classic to new genomic tools: an example with interspecific tetraploid Urochloa spp. hybrids

• Ano de defesa: 2018
• Autor(a) principal: Matias, Filipe Inácio
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Brachiaria spp.; Allele dosage; Associação genômica; Dosagem alélica; Genomic association; Genomic selection; Genotipagem por sequenciamento; Genotyping-by-sequencing; Mixed models; Modelos mistos; Poliploidia; Seleção genômica
• Link de acesso: http://www.teses.usp.br/teses/disponiveis/11/11137/tde-21032019-153059/

A tropical forage breeding program contains several peculiarities, especially when it involves polyploid species and facultative apomixis. Despite their importance, there is still a lack of information on genetic studies of critical forage traits and on the employment of genomic tools when compared to other crops and temperate forages. The genus Brachiaria is the most important for forage in tropical regions mainly beef production. The commercial species in this genus are excellent perennial forage, and the identification of superior genotypes depends on the selection of many characteristics under complex genetic control, with high cost and time-consuming evaluation. Therefore, the knowledge about uses and applications of classic and genomic tools in forage traits may be useful to support breeding programs and the development of new cultivars. In this context, the aim was to evaluate several different classic and genomic tools to be employed as selection strategies in a traditional tropical forage breeding program. A panel of tetraploid hybrids obtained from crossing Urochloa brizantha x Urochloa ruziziensis was phenotyped and genotyped to evaluate genetic parameters and perform genomic studies. The classic phenotypic analysis showed no clear trend of the importance of additive and non-additive genetics effects for agronomical and nutritional traits. The Mulamba and Mock index should be used in the univariate level, due to the promotion of a more balanced response to selection for all traits in the multivariate selection. In the genomic extraction and evaluations, the reads that were aligned to a \'mock\' reference genome, created from GBS data of the cultivar \'Marandu\', had more SNP discovered compared to the closest true reference genomes, Setaria viridis and S. italica. We recommended different thresholds of sample depth and genotype quality (GQ) to eliminate poor quality reads without introducing genotype bias. Cross-validation revealed that missing genotypes were imputed with a median accuracy of 0.85 using Random Forest algorithm to produce a complete genotype matrix, regardless of heterozygote frequency. The genome-wide association analysis (GWAS) revealed candidate genes associated with many tropical forage traits across all cutting seasons, which could be the first step toward marker-assisted selection (MAS). Moreover, our results suggest that accounting for allele dosage is essential, since the tetraploid level provided more information about the true biological state. Therefore, our findings revealed the complexity of the genetic architecture of Urochloa spp. traits and provided important insights towards the application of GWAS in polyploids species. The genomic selection analysis revealed that GBLUP-A (additive) and GBLUP-AD (additive + dominance) showed similar prediction abilities considering both single and multi-trait models. Conversely, combining GBLUP-AD and tetraploid information could improve the selection coincidence. Furthermore, the multi-trait validation scheme 2 (VS2), where one trait is not evaluated for some individuals, provided an increment of up to 30% to the prediction ability. Therefore, it is an useful strategy for traits with low heritability. Overall, all genomic selection models considered provided greater genetic gains than the phenotypic selection. Similarly, the allele dosage associated with additive, dominance and multi-trait factors increased the accuracy of genomic prediction models for interspecific polyploid hybrids. Finally, genomic tools should be used in forages breeding programs in order to reduce cost and time.